NEW UPDATED Room Calibration for Film and TV Post

[Paul Maunder]

I understand all the room calibration stuff, but I have a question regarding how digital mix levels translate to the analog doman. If I produce a mix with an RMS level of -20dBFS which corresponds to OVU, what happens to the absolute peaks in the program (which could be up to 20dB higher than the RMS level). Won't they clip if I'm putting the mix down to an analog tape format (for example)?

[Tiago Silva]

I understand all the room calibration stuff, but I have a question regarding how digital mix levels translate to the analog doman. If I produce a mix with an RMS level of -20dBFS which corresponds to OVU, what happens to the absolute peaks in the program (which could be up to 20dB higher than the RMS level). Won't they clip if I'm putting the mix down to an analog tape format (for example)?

Yes they will. That's why the usual broadcast level recommendation is to limit the peak levels to 10dB over the 0VU digital level chosen as reference: 0VU= -20dBFS, peak level= -10dBFS; 0VU= -18dBFS, peak level= -8dBFS, etc. This way, when you align the analog machines at the chosen 0VU, you'll notice the max level in the VU meters won't go very far above +3VU, and won't clip unless you're really hammering the peaks with an L1.

[RichE]

Hi All,

I would like to thank you all for this thread.... I am just getting a room setup and your comments, links and knowledge are great., invaluable!
For you interest, and in reply to an earlier piece about the music guys.... I have been a recording engineer freelancing for 15 years, during which time I had the great pleasure of assisting with Michael Kamen on a number of film soundtracks and mixing others. My point isn't about room setup - but how it translates.In a room pre-referenced by the Dolby guys, we used to mix to a dialogue track that we believed had been recorded to DAT at a ref level ... normally a 1khz tone at "0"VU = -15 to 18 db on the DAT meter. We would then mix the music to that audio level ... most of the time the music would just be loud out of choice:) So it would end up close to the 0db level on the DAT.
At that time it wasn't about destroying dynamics with compression like now, but is was amazing how many times we went to the final dub and the music was so much quieter than expected
Having read this thread I can now understand why.... if I was you all and I put on a DAT music track and suddenly it was close to 20db louder than I was used to listening to then I would think we were all nuts.... you have to remember that our actual volume in the studio is up and down all day long depending on which instrument you listen to.
Hope this makes some sense of it all.... roll on the AES with a proper standard, cause there are a lot of music engineers coming your way in to 5.1
Thanks again.
RichE

[Chief Technician]

The following is some Q&A about the dialnorm parameter in Dolby Digital/AC-3 from a newsgroup I'm on. I thought some people might find it useful.

Question:
Is there something like dialnorm in the Dolby Cinema processors ?
Dolby rep (name removed) - No there is no dialnorm in 35mm film because
the Dolby Digital system used for 35mm film does not include metadata.
This is because all cinemas that play this soundtrack have 5.1 channel
loudspeakers (so you don't need downmixing) which are capable of wide
dynamic ranges (so you don't need dynamic range compression) and are
set at a calibrated replay level matching that in the dubbing studio
(so you don't need dialogue level).

Of course, the very latest version of our cinema processor also decodes
the Dolby Digital stream that you find on a DVD or a TV broadcast and
this does use metadata - in this case we can use the dialnorm value but
we do not downmix or change the dynamic range for the reasons listed
above.

Question:
Am I correct in stating that when you leave dialnorm at -31dBFS
and set the dyn rng preset to "None", nothing is done to your audio?

Dolby rep (name removed) - No. When you create a 5.1 Dolby Digital bitstream for
TV or DVD it is possible that this bitstream will be decoded as a stereo
or mono downmix. Because downmixing 5.1 channels to 2 or 1 channel is an
additive process it is possible that, unless protected, the signal will
exceed 0dBFS. The dynamic range control metadata is used to prevent
clipping in signals that would have exceeded 0dBFS. Therefore we make
two calculations when working out the dynamic range metadata - the first
is the value that you want based on the preset selected, such as film
standard; the second is a value that prevents clipping when the signal
is downmixed. So if you set the preset to None then the metadata will
only be values to prevent clipping. These values depend on the level
of your 5.1 mix and the dialnorm value (a dialnorm of -20 for example
means 11dB of attenuation so 11dB more headroom when downmixing to
stereo and less likelihood of exceeding 0dBFS).

One common mistake is for music mixes that are compressed and normalised
like a CD to have their dialnorm values at -31dB and the DRC at None.
While this sounds no different in 5.1, in stereo the music is compressed
by up to 11dB in line mode to prevent signal clipping. Setting the
dialnorm at -20dB (where it should be) removes this dynamic range
metadata completely.

Of course when listening in 5.1 you can always listen to the full
dynamic range irrespective of whether the DRC preset is None or Film
Standard or anything else.

[Sound Guy Andy]

Funkcity said:

It is for film:
-20 dBFS = 0VU = +4 dBu = 85 dBC SPL for L,C,R, 82 dBC SPL for LS, RS and on a Real Time Analyzer 10 dB of inband gain above the center channel from 20 ~ 30 hz to 120 hz

It is for TV, DVD:
-20 dBFS = 0VU = +4 dBu = 85 dBC SPL for L,C,R,LS,RS and on a Real Time Analyzer 10 dB of inband gain above the center channel from 20 ~ 30 hz to 120 hz

Are you referring to the Lfe channel when you say, "10 dB of inband gain above the center channel from 20 ~ 30 hz to 120 hz?" If so, I understand that we monitor the Lfe 10dB louder than other channels (and therefore print it 10dB lower, since its given 10dB of extra gain upon playback), but why do we calibrate the Lfe over the center channel?

I have some other questions about print levels for downmixes, but I'll start new threads for those.

Thanks to everyone for all this great info!


Andre

[Aquarius]

Hi Georgia,

I'm using a DSM7.1 too but I'm not really confident in calibrating my room with it....
What are the results of your testing against Dolby PT session ?
You posted you could send a Dolby reference session, would it be possible if I leave my mail ? ([email protected])

Best regards,
O.

[funkcity]

<<>>>


Yes
But you must observe it on a Real Time Analyser NOT an SPL meter.
This has been the standard as long as I can remember.

[funkcity]

Hey Marty

How long will this thread continue??

[Ciaran]

Having been an avid reader of this thread for quiet some time, I thought of sharing a question with you regarding "session calibration" - in the sense of calibrating level meters in a PT session to accurately read track RMS and Peak Levels.

I have read the posted solutions regarding the BFactory Meter issue - namely, bus a +3dB signal from the track being monitored to a track with the BF Meter on it, and then a -20dB test sine tone will read 0Vu on the BF Meter on its RMS setting. Is the idea of this to make -20dB=0VU simply for the sake of conveniant metering?

In my tests today, I firstly sent a Blue Sky sine wave at 1kHz and -20dB through the BF Meter and it approximately read -3.1Vu. A sine wave at 1kHz and -20dB from the DIGI signal generator plugin through the same BF Meter reads -3Vu.

However, I recently got the Waves bundle and have been trying to use the Waves PAZ meter for the same purpose. I've noticed that sending the Blue Sky through PAZ(setting on RMS) results in an RMS level of -23.9dB (-3.9VU) with either no weighting or a B or C weighting and PAZ reads -23.1dB (-3.1VU) with an A weighting.

A sine wave at 1kHz and -20dB from the DIGI signal generator plugin through PAZ reads -3.4Vu or -23.4dB.

I also noted that by generating pink noise with the same generator, a -20.5dB or -.5Vu level results on the PAZ and an unsteady -1Vu on the BF Meter.

Can anyone explain what is going on, and why a generator of a pure sine tone at -20dB is only giving an RMS level of -23dB/-3Vu? If this is the way it is supposed to be, grand, but I am confused. Assuming that we calculate the average power by [RMS=Vmax/root2] I guess this is expected right?

If that is the case, the software system itself is accurately calibrated to -20dB rather than the 0VU RMS level, which makes sense, but I guess it is making me confused as to the capacity or the need to monitor a mix at 0VU=-20dB when they don't seem to match.

Hopefully you can explain my ignorant argument on the solution as right now, I am all washed out with numbers in my head and need to switch off my brain for the night!


yours in music

Ciaran
www.ciaranhope.com

[dr sound]

Poster: funkcity
Subject: Re: UPDATED Room Calibration for Film and TV Post

Hey Marty

How long will this thread continue??

dr. sound replies;
It will continue until you learn to spell my name correctly!